A systematic review of adverse events in randomized trials assessing immune checkpoint inhibitors

The advent of immune checkpoint-inhibitors (CPI) has transformed treatment for several cancer types. This review was performed to assess the rate of adverse events (AEs) associated with the use of CPI, alone or in combinations. A review of AEs reporting quality was also performed. All publications of Randomized Clinical Trials (RCTs) assessing CPI published before December 2017 were included. To investigate the quality of AEs reporting, a set of items was defined based on the 2004 CONSORT harms extension statement. Rates of Grade 5, serious, and study-withdrawal related AEs were collected in each treatment category. Specific immune related AEs (irAEs) were also collected when available. Pooled estimates of adverse event rates were calculated by using generalized linear mixed model. A total of 35 RCTs including 16,485 patients were included. The overall quality of AEs reporting was satisfactory, but items pertaining to methods of data collection and analysis were infrequently reported. Grade ≥ 3 AEs were reported for 14% (95% CI 12-16) of patients treated with PD(L)-1 inhibitors, 34% (95% CI 27-42) of patients treated with CTLA-4 inhibitors, 55% (95% CI 51-59) of patients on CPI combinations and 46% (95% CI 40-53) of patients on immunotherapy-chemotherapy combination. The profile of irAEs was different among the treatment categories. The use of CPI, especially in combination, is associated with significant rates of Grade ≥ 3 AEs. Healthcare planning should anticipate the expected high number of patients presenting with irAEs in the future.